ashishbora / csgm Goto Github PK

hello，when i was running your code I could not find the “mnist_utils“，could you tell me that how could i find it？

Hi, I am interested in replicating your work for my domain area. Could you point out the code where the generator is frozen, and the latent vector is optimized? Thank you.

The pretrained model files are available here
https://www.dropbox.com/s/3o2vi1w1wde0ids/bora-pretrained.zip?dl=0

Could you please provide a working download link for datasets and models!

Thank you in advance!

Best regards!

here is the problem, when I run the demos, each returns
Traceback (most recent call last):
File "./src/compressed_sensing.py", line 177, in
main(HPARAMS)
File "./src/compressed_sensing.py", line 19, in main
xs_dict = model_input(hparams)
File "/home/feiht/csgm/src/mnist_input.py", line 56, in model_input
mnist = input_data.read_data_sets('./data/mnist', one_hot=True)
File "/usr/local/lib/python2.7/dist-packages/tensorflow/python/util/deprecation.py", line 306, in new_func
return func(*args, **kwargs)
File "/usr/local/lib/python2.7/dist-packages/tensorflow/contrib/learn/python/learn/datasets/mnist.py", line 262, in read_data_sets
train_images = extract_images(f)
File "/usr/local/lib/python2.7/dist-packages/tensorflow/python/util/deprecation.py", line 306, in new_func
return func(*args, **kwargs)
File "/usr/local/lib/python2.7/dist-packages/tensorflow/contrib/learn/python/learn/datasets/mnist.py", line 62, in extract_images
magic = _read32(bytestream)
File "/usr/local/lib/python2.7/dist-packages/tensorflow/contrib/learn/python/learn/datasets/mnist.py", line 43, in _read32
return numpy.frombuffer(bytestream.read(4), dtype=dt)[0]
IndexError: index 0 is out of bounds for axis 0 with size 0

why is it, thank you!

Hi Ashish

I'm wondering how to train the dcgan on another dataset? I have tried using code from https://github.com/carpedm20/DCGAN-tensorflow to train the model and then directly use it in this project but it doesn't work

I don't understand why you use ||AG(z)-y|| as the loss function instead of using ||G(z)-x|| which makes more sense to me. I intuitively feel that multiplying A with x will lose lots of information. Looking forward to your explanation. Thanks!

hello,I want to consult with about your loss function for vae and GAN. In your paper , the loss function is ||AG(Z) - y||^2,but you used 'tf.matmul(x_hat_batch, A, name='y_hat_batch') ' to finish the funtion, but this code means ||G(Z)A - y|| ^ 2, there is no influence for the fuction,isn't it? I really hope to hear from you soon. Thank you very much!

In the dcgan_estimator(hparams) function of celebA_estimators.py, shouldn't the global_step variable and other variables of AdamOptimizer get reinitialized in each call of the nested estimator(A_val, y_batch_val, hparams) function?

The file includes " python ./src/main.py", whereas "main.py" is not found in file "src". Would you mind assisting me with this question? thank you very much.

I would like to replicate your results, but the API of some of the dependencies has changed, thus some of the examples do not work. It would help a lot if you could post a version number for each of the dependencies, where all examples are guaranteed to run as expected.